Google has a surprise in store for all of us who use its Artificial Intelligence tools: the AI assistant, Google Bard, has been updated worldwide with Gemini Pro. Not only that, but a tool we’ve been waiting for for a long time has also won: we can finally generate any image by asking Google for it.
A dog riding a scooter down a Bombay street, a cat putting nails on the Statue of Liberty and any other crazy thing you can think of: the magic of generative AI turns wishes into reality. This function has revolutionized creativity, it has also created controversies almost since its appearance. And little by little, renowned technology companies, like Google, are joining the AI wave. The bard levels up.
Google Bard is tied with ChatGPT on Bing
With the February update, Google Bard received a new language model worldwide, Gemini Pro. This model promises a better understanding of commands, Bard’s creativity increased, its power and the variety of results. On the other hand, the wizard hosted in a search engine area now also generates images.
The novelty of AI-generated photos is not available worldwide: Google has decided to restrict it in the United States, something quite common in popular versions. Additionally, there is another limitation to consider: we need to speak to the assistant in english. Just to ask for the images, the rest of the time it supports Spanish.
With the two premises clearly stated, limitation to the United States and commands in English, let’s see how Google Bard image generation can be tested anywhere in the world:
- Open your VPN app and select a US server. We used NordVPN, but any will do. Like TunnelBear.
- Load the Bard page, you will see that it appears in Spanish. If you view the side menu, “United States” should appear in the bottom location.
- Write the description of the image you want to generate. Commands for creating graphic content They must be in English.
- If you don’t know how to translate what you want, you can use Google Translate or Bard itself: write the command in Spanish, ask it to translate it for you, and copy/paste the command in English.
- Google Bard will offer you two samples based on what you requested. If you’re not convinced, click “Generate More”. You can also refine the order by repeating the request with more emphasis on a detail, adding style, or removing adjectives. A good image usually doesn’t come out the first time.
There is a geographic limitation, but you can get around it with a simple VPN connected to US servers. The generation is available both from the computer and from the mobile: just load Google Bard in the browser and that’s it. According to our tests, the level of creations is quite high and without too many artifacts so typical of AI.
Cover image | DALL-E 3 in ChatGPT edited
In Xataka Android | AI duel: I faced ChatGPT against Google Bard and I clearly know who is the absolute winner